[FlorianO]

[I posted this on DIYaudio full range forum but was redirected here as more specialized forum. Sorry if you already seen this. 10x.]

I have a bit of a problem understanding the differences btw using one or two woofers + FR on flat OBs and the related baffle size.

What I am looking at is the following:

- Visaton NoBox and its variants by Dick Olsher: Single 15 in woofer (Hawthorne Audio Auggie, Viaston BSG40) + FR (B200, Fostex 208 Sigma, Feastrex D5nf, Lowther DX55 and C55) on a relatively narrow baffle (aprox 60 cm), slightly sloped.

- BD-Design Quasar MkII. Even narrower baffle (46 cm) but much thicker and straight. Two 15 inch woofers (Bert's own BD15 15 inch woofers, other ppl employing the Eminence 15A) + FR (Fostex FE206E, AER MkI).

- Martin J King projects with same dual Eminence 15A + FR (Lowther PMA2A, Fostex FE 167, ) but on super-large baffle (150 cm wide).

My questions:

- What is the difference btw employing one 15 incher vs two in an OB ? When does one go for one and when for two ?  Any baffle size related issues (I am thinking very narrow Quasar with two woofers vs Martin J King OB project with same two woofers but 150 cm wide baffles)


- Very wide baffle, narrow baffle, even narrower but thick baffle ? Any reason for going for one of the other while employing similar (or the very same) woofers ? Needless to say I take for granted that both Dick, Martin and Bert know what they're doing so it must be a matter of compromises -- in that case, what are those ?

Thanks in advance for shedding some light into this,

Florian

[JohninCR]

Florian,

Double the baffle (for flat baffles), moves the dipole roll-off point lower by 1 octave.  This means +6db bottom end potential.  Adding another driver also gives you +6db at the bottom, but doesn't change where dipole cancellation starts. 

Your design goals will dictate everything.  eg MJK was trying to match the sensitivity of his lowthers AND minimize bass correction (he uses only +2db of bass boost), so he was pretty much forced into using all of the tricks.... dual high Q woofers + big baffle + U type extension behind the woofers.

If you accept that you'll use EQ to flatten response and separately power your bass augmentation, then design becomes much easier.  Just use Linkwitz's DipoleSPLmax spreadsheet and you can determine the bottom end potential with the woofer(s) and baffle you can live with.  It also enables you to use narrower baffles around your fullranger, which I have found to lead to better imaging.  The trade-off as go with small baffles around the fullranger is that tonality changes due to the smaller area of 1/2 space wavelaunch.

Curiously, you have left out the 2 primary "scientists" regarding how to design open baffle alignments, Siegfried Linkwitz and John Kreskovsky, who have generously shared a wealth of invaluable info on their websites www.LinkwitzLab.com and www.MusicAndDesign.com .

FWIW I consider all of the designs you mentioned to be quite compromised and demonstrates that even very experienced designers of box speakers don't yet fully understand how open alignments function.  2 years ago they were still ignoring OB's, so it's good that they're trying now.  Of those mentioned, I like the Quasar best, although I question the use of the "biting" FE206, along with the lack of consideration of pipe resonances from the cavity behind the main driver and edge diffraction.  Just minor changes could make it outstanding and quite room friendly design.

[MJK]

If you accept that you'll use EQ to flatten response and separately power your bass augmentation, then design becomes much easier.  Just use Linkwitz's DipoleSPLmax spreadsheet and you can determine the bottom end potential with the woofer(s) and baffle you can live with.  It also enables you to use narrower baffles around your fullranger, which I have found to lead to better imaging.  The trade-off as go with small baffles around the fullranger is that tonality changes due to the smaller area of 1/2 space wavelaunch.

You are also going to need to push the drivers harder in a narrow baffle. More power from the amp and higher displacements. It requires EQ which probably almost locks you into an active bi-amped setup. The low end drivers and amps need to work much harder in a narrow baffle design to produce 40 Hz.

FWIW I consider all of the designs you mentioned to be quite compromised and demonstrates that even very experienced designers of box speakers don't yet fully understand how open alignments function.

How about listing the compromises for each design and what you would do differently? What are we all missing that you fully understand?

[JohninCR]

I have no problem having that kind of discussion if it's 2-way and you drop the condescending tone.  I don't claim to have all the answers, but I do believe there's more to it than mounting drivers on a big piece of wood.  I'll kick the discussion off by saying that I believe 40sqft of baffle area for a pair of indoor speakers is a compromise that very few will be willing to make.  In addition, I found that such large baffles also have sonic compromises.

[FlorianO]

John, Martin,

Thank you very much for chipping in.  I knew about the Linkwitz site and his setup with bipolar woofers. I didn't knew about the other, will take a look.

In my beginner "purist" mind I do not like the idea of electronically EQing the woofer response but using a plate amp for driving the woofers is  smth I would very happily consider and  live with, but with a (very very big) caveat: that the whole thing integrates properly  (which was _not_ the case when I listened to the Bastanis Prometheus MkII).

Again, I take for granted that both Martin and Dick really do know their stuff  but both were using high Qt, low fs 15 inchers (Eminence 15A and resp. Augie) on very different baffles. Martin mentions explicitely on his site that on the very large OB the driver excursion was minimal.  If the added benefit of using such a  very wide baffle is limited driver excursion I would correlate that with faster response/better dynamics (pls correct me if not).  Same thing with smaller Qt drivers (?)

Many thanks for the information and pointers. Again, what I am after here is _understanding_  the whys and the hows.

Thanks again,

Florian

P.S. Forgot to mention the Hawthorne Audio Duet from the list of flat OBs above. Anyway.

[FlorianO]

FWIW I consider all of the designs you mentioned to be quite compromised

Any other dipole woofer(s) + FR on flat OB design I should look at ? Or was that the dipole woofer + FR "concept" that is compromised ?

Thanks,

Florian

[maxro]

FWIW I consider all of the designs you mentioned to be quite compromised and demonstrates that even very experienced designers of box speakers don't yet fully understand how open alignments function.

How about listing the compromises for each design and what you would do differently? What are we all missing that you fully understand?

I doubt anyone "fully" understands open baffle behaviour. That said, John, MJK's question as to what you would do differently is a valid one. Further, having that discussion privately will not help the OB (r)evolution. Ought you not share your collective knowledge with the DIY masses? 

[JohninCR]

Max,

I wasn't looking for a private discussion, just trying to make it 2-way instead of me spouting off opinions that are nothing I haven't posted before.


Florian,

Many OB designs would apply.  Just replace the mid + tweeter with a full ranger.  The trickiest part of your plan will likely be the FR/woofer(s) integration because some type of XO is required and the fixed slope low pass filter of a subwoofer plate amp may prove inadequate due to the slope and/or effect on phase.  As Martin mentioned the smaller the baffle the more air the woofer(s) must move, but this isn't necessarily a bad thing as long as you aren't over-working them.  That's why I said that you should start with your performance goals and the physical size you are willing to accept.  There are even ways to flatten bass without EQ.  Placement can help, along with driver selection and the physical construction itself.

If you want to be a "purist", there's also the option of seeing how far you can take a fullranger by itself.  I'm going to be eating some crow very soon because I insisted on numerous occasions that the B200 couldn't go it alone unless you listen only to chamber music, yet I've been happily listening to a pair for a few days without EQ and no bass augmentation at all.  I'm using only a simple filter to subtract 3db centered about 2khz to tame the hot top end, and a steep rumble filter at 40hz that has a high Q so it adds a little boost above 50hz.  The results are surprising me so much that I want to get measured results before posting all the details.

[MJK]

I have no problem having that kind of discussion if it's 2-way and you drop the condescending tone.

John,

You made a statement and all I did was ask for some further explanation. Nothing condescending intended by asking a question.

instead of me spouting off opinions

I was hoping for technically correct facts not opinions, opinions are just a statement of why you made the compromises/choices that lead to your particular style of design. Different people make different compromises and produce different designs to meet their design goals. Each has strengths and weaknesses. I have pretty much stopped posting on DIY forums because people tend to mistake opinions and speculation for technically correct physics based facts and then want to argue the points. OK, if you don't want to discuss technically defendable facts as to why you feel we are all misunderstanding OB design I will move on to other things. No big deal.

Martin

[FlorianO]

Martin, John,

As I stated already, what I am after is understanding the _physics_  of it all (the fact that I am a complete beginner in understanding the physics of flat OBs and how driver parameters affects their performance on OBs would make that journey longer, but still, I am working hard to catch up).

What I would readily accept is that "there are more ways to skin a cat". IOW it is all a matter of compromises and each has their own preferences.

Very different than taking oppinions as "facts".

Since the hybrid design (one or two woofer(s) + one FR) seems to come more and more often I limited my question to this particular setup and was trying to figure out the technical compromises of the different solutions (as above).

Again, it's all physics and (at least I am) on a path to better understanding.

Thanks all for shedding light into this. For sure some of the material  was covered elsewhere (am I working my way through the Linkwitz site atm) but different posting on various forums confused me quite a lot...

Sincere regards,

Florian

[MJK]

Florian,

What I would readily accept is that "there are more ways to skin a cat". IOW it is all a matter of compromises and each has their own preferences.

Exactly. A great way to study the trade-offs between baffle size, driver selection, low or high Qts and fs, crossover selection, low frequency boost, and the resulting SPL and displacement response curves is to run a study using computer simulations. In an afternoon you can learn a lot about the different trade-offs and decide for yourself what will be required to produce your desired low frequency response. Granted the simulation results will not tell you exactly how a finished speaker will sound but it will tell you how low and how balanced the SPL response will be when you build the design. Simulations will eliminate many of the trial and error mistakes. I believe running simulation is much more efficient, insightful, and produce a deeper understanding of speaker behavior and trade-offs compared to building a series of experimental prototypes and guessing at what is creating the good and bad aspects of the final SPL results. If used correctly, computer modeling and measurements are short-cuts to an optimized design in only a couple of hours.

Martin

[Davey]

 I have pretty much stopped posting on DIY forums because people tend to mistake opinions and speculation for technically correct physics based facts and then want to argue the points. OK, if you don't want to discuss technically defendable facts as to why you feel we are all misunderstanding OB design I will move on to other things. No big deal.

Martin

An excellent comment, and exactly correct.  If there's one thing I've noticed through the years monitoring audio forums its that a persons posting frequency and their knowledge is an inverse relationship.  Generally. 
Of course, there are exceptions like me who don't post much and don't know much either. 

Cheers,

Davey.

[johnk...]

 I have pretty much stopped posting on DIY forums because people tend to mistake opinions and speculation for technically correct physics based facts and then want to argue the points. OK, if you don't want to discuss technically defendable facts as to why you feel we are all misunderstanding OB design I will move on to other things. No big deal.

Martin

An excellent comment, and exactly correct.  If there's one thing I've noticed through the years monitoring audio forums its that a persons posting frequency and their knowledge is an inverse relationship. 
Of course, there are exceptions like me who don't post much and don't know much either. 

Cheers,

Davey.

I think I have to agree with Martin and you too Davey on this. I try to stay active in some DIY forums, but it tries my patients at time. Of course, I have to admit to besing somewhat of a lightening rod and verbal rock thrower. But you know what they say about "words". 

[JohninCR]

Martin,

I think part of the problem stems from taking things differently than they are meant.  I didn't say "all" or "misunderstanding".

One flaw I see in most designs is the lack of consideration for how the rear wave expands.  The rear wave is part of the primary response, yet we see many people construct things on the back side of their baffle that they wouldn't even consider putting on the front.  These structures become even more of an issue once you get a wall reasonably close behind them, especially the larger ones that are typically somewhat flimsy in construction.  The result is misdirected reflections, panel vibrations, and cavity resonances.

I consider these things factual, not opinion.  If you disagree, then I suggest you give the math a break and spend some time experimenting.  Mount one of your Lowthers backward and turn that baffle around and give it a listen.  Play some bass rich music at moderately loud levels and go put your hand on different parts of your construction.  Cavity resonance may not be the right terminology, so I'll just call it a hollow sound.  Just because you crunched some numbers that said resonance won't occur, doesn't mean you crunched the right numbers.  Having the rear wave of your fullrangers playing out from between 2 parallel panels must cause problems.  Also, with your placement the entire structure forms a large cavity between the walls, equipment rack, and baffles which most likely affects the lower bass.

Opinion stuff would be things like, I don't like the way large baffles block such a large chunk of the rear wave, that I haven't liked any of my test baffles with the shortest distance being over the top, that I think 4'x5' baffles are plain silly, or that Visaton's NoBox was just an off the cuff design by an engineer who believed in boxes but the surprisingly good sound made them not even try to optimize it in any way (I've seen similarly bad OB's suggested by Supravox).

Don't get me wrong, I believe wholeheartedly that your speakers sound great, but I want to encourage people to come up with designs that are even better both in form and theory.  It wasn't much more than a year ago that most OB designs were based on a popular science article from the 60's.


Davey, John, and Martin -  If you're referring to me, since I'm one of the most active posters in this forum, feel free to correct me when I'm wrong.  It's part of the learning process and I have thick skin.  Just understand that a condescending attitude is likely to reflect right back if I'm not.  Davey, we're still waiting for the measurements you promised months ago (I think it was in the topless U thread).

[Davey]

John,

I didn't forget, but thanks for reminding me.  This is still on my to-do list but near the bottom.  Other, car-related, projects have been taking up my free time lately.

I pretty much know what the plots will show and I would be just re-tracing the footsteps of more qualified designers than myself who have already taken such measurements, but nevertheless it's still a good exercise and gives me more experience with my testing procedures.

Cheers,

Davey.

[9anda1f]

A little of my own personal perspective...

Basic scientific method says:
1) observe
2) postulate/hypothesize
3) experiment
4) evaluate the results and adjust perspective for further observation - repeat the cycle

Traditional practice designs physical experiments based on well-thought-out hypothesis...the cost of preparing an experiment is variable and based on materials and time expended.  Physical modeling is often involved prior to actual experiment conduct.  Context is a critical variable.

Modern practice uses a process of computer/mathmatical modeling to help minimize costs associated with extensive physical experimentation.  There is a cycle involved in mathematical modeling that includes design, verification, and validation (and accreditation too sometimes).  Verification is the process that determines whether the model performs as designed.  Validation is the process that determines whether the model accurately reflects real-world performance.  (Accreditation is the process that determines if a model is a standard by which to judge real-world performance in multi-variable conditions).  The cost of mathematical modeling is variable and based on time expended to build/maintain the model.  Models are exercised within a simulation that provides the context over time.

Model validation involves experimentation/testing in the real-world.  Measurement/data collection is required, and either ascertains that the model (in the context of the experiment) is accurate or feeds back correction information into the iterative design process to adjust the model to reflect observation. 

Overall, what I've called "traditional" and "modern" applications of scientific method are complementary approaches to exploring "reality".  The least interesting outcome of these two approaches is when all results of experimentation agree.  The MOST interesting outcome is when the results of these two approaches to understanding DON'T agree (or appear not to agree). 

When results from two (or more) different approaches to understanding phenomena appear to disagree we can hypothesize that:
1)  there is a difference in semantics (i.e., we are not talking the same language and so our results merely APPEAR to disagree)
2)  there are unaccounted variables at work in the simulation affecting the outcomes
3)  our model is flawed (be it physical or mathematical)

So, all of this to say that I am MOST intrigued and interested when John's and Martin's views DON'T agree!  These are the times when the most interesting dialogue occurs, the thinking caps go on, and progress is made!  Pursuing truth from multiple approaches is a self-correcting process. 

...two cents, for what it's worth!   

9anda1f

[JohninCR]

9anda1f,

I much prefer when I get results that agree with the theoretical, although I sure would like to find a free way to split water into H and O2.   I don't think Martin and I really disagree anyway.  It's probably only a semantics problem.  I just threw the idea out there that the commercial box design experts starting to dabble in OB's have new things to discover and learn, and I wasn't really including Martin in my thought process.  I'm sorry if my choice of words was too strong.

I've built enough baffles with shapes similar to open backed boxes to know that they introduce new problem issues, so when I see these shapes with nothing done to address those issues I speak up.  I'll leave it to guys like Martin to define and quantify the issues, and restrict myself to trying to figure out ways around them in order to take advantage of the benefits of folding a baffle.

[FlorianO]

Florian,

Exactly. A great way to study the trade-offs between baffle size, driver selection, low or high Qts and fs, crossover selection, low frequency boost, and the resulting SPL and displacement response curves is to run a study using computer simulations. In an afternoon you can learn a lot about the different trade-offs and decide for yourself what will be required to produce your desired low frequency response.

As a former researcher I do hold simulations very, very near&dear. But (this is an old pet peeve of mine, and old habits die hard  ) "empiric" simulations are not exactly my fav.  Generally I do my best to first master the theory (or at least have a grasp  of the main issues) before delving into simulations. IOW know what the h@#ll I am doing before simulating:).

 That is exactly why my intention with this thread was to instill that kind of discussion that brings those technical issues forward:
- woofers with low vs high Qts and related compromises ("are woofers with low Qt `faster`  and `tighter` but need more excursion ?" why the one and not the other).
- woofer fs influence. same with xmax
- baffle width, baffle diffraction step effects and how does it affect the above
- compromises in woofer/driver design (high Qt and low fs results in low sensitivity)
- problems with using one vs two woofers, bi-amping or not
- etc, etc, etc.

...and all that in a very specific context: flat OB (narrow or very wide) with woofer(s) + FR. no more, no less.

I'm sure that for expert designers like you and others on this forum this is already boring and  was already covered, maybe several times over. While I can grok the physics (in due time, at least..) I am almost sure that the gory details and small compromises that need to be done in real life are a black magic avail only to the very experienced few...

Granted the simulation results will not tell you exactly how a finished speaker will sound but it will tell you how low and how balanced the SPL response will be when you build the design. Simulations will eliminate many of the trial and error mistakes.

Been there, done that (albeit not with speaker design) so I understand perfectly the message. However, properly correlating simulations with real world example tends to be a skill only developed through experience. Which,  not surprisingly  (to me at least) is exactly what John is saying . All in all we all want to get there, but choose different paths  -- and sharing the accumulated knowledge can only benefit everybody.

Sorry for the lhorribly OT rant.... I just wanted to highlight (again) that what I am after is understanding the theory _and_ be able to strongly corelate that with net results and real life issues and compromises. And all that just in order to get my bearings straight .

Once again, many thanks for the fruitful technical discussion and pointers to relevant material/resources,

Florian
 

[AJinFLA]

I try to stay active in some DIY forums, but it tries my patients at time.

I thought your patients were the DIY folks. Hmm

Davey, we're still waiting for the measurements you promised months ago

John, I still don't see a single solitary measurement from you. The OB/RLH/worlds only,etc,etc? That tapered H baffle mid (I think you called it a waveguide mid)? What ever happened to proving all the measurement geeks wrong? What about your promise to me?

I much prefer when I get results that agree with the theoretical

I still have no idea how you do that without measuring

C'mon John, point a mic at one of those contraptions of yours so we can see what's really going on 

cheers,

AJ

[JohninCR]

Florian,

You aren't ranting or OT either.  I like playing around with simulations and have lost myself in the world of simulations for days using Martin's spreadsheets and HornResp, however, I haven't bothered to sim a single OB.  That doesn't mean I employ a trial and error approach, quite the contrary.  I just feel the OB's are to simple in some respects and too complicated in others for simulations to be of use, at least for me especially due to the complex structures I am exploring.  As I move out of construction mode and into measurements, I plan to get all the low priced simulation programs available and see how well they predict reality, since those who post sims seldom post real world measurements.  I'm encouraged by Martin's newest version, which I understand includes the effect of the front wall in the overall equation.

While I'm still on the topic of simulations, I have a couple of questions.  Where are the real world measurements that support their meaningfulness?  In Martin's case, how was the effect of that 12" U type cavity factored into the simulation?  Also, in his discussions about his project, the simulations mention plots of expected response at 1m.  Is that really true, because to have a nearly flat listening position response an accurate 1m simulation would need to reflect an 8-10db higher bass response at 1m???

Linkwitz provides the only 2 tools I use or need, and they apply to all open alignments.  It takes me about a minute to get the info I need.    Add less than an hour to build some test structure to hold the drivers.  Then in the case flat baffles just add a roll of duct tape, a razor knife, cardboard, and plenty of beer, and that makes for a pleasant remainder of the afternoon listening to music using different versions of reality and evaluating the real world effect of baffle and/or placement changes.

The things I use are the formula for Fequal, and as I mentioned before, his DipoleSPLmax spreadsheet.  Fequal is the frequency where output is equal to monopole.  In other words, where you have delayed the rear wave by a distance that makes it 90deg out of phase with the front wave so it neither cancels nor reinforces the front wave.  That formula boils down to Fequal = 58.3hz / D in meters, where D is the Difference in distance that the front and rear waves must travel to reach your listening position.  For a flat baffle D is essentially the center of the driver to the baffle edge.  Below Fequal dipole cancellation of -6db/oct kicks in, and above that point (assuming a circular baffle in freespace) SPL climbs to +6db at Fmax where the front and rear waves are directly in phase, so it's like adding another driver.  Fmax occurs at about 3X Fequal.  This increase isn't linear and works out to about +4db in the first octave.  Then an octave above that (Fmax X 2), is the first dipole null.

Peaks and nulls above Fequal is where off-center mounting helps smooth response by creating different travel distances and averaging out the peaks and nulls for a smoother response, because they occur at each octave higher.  In the bass frequencies the wavelengths are so long that for woofers the deep null is difficult to avoid, and must be considered in your XO strategy.

Yes the floor boundary is useful to increase bass response, and yes the front wall boundary can make careful placement offset a portion of the dipole cancellation, but in reality there's no getting around the sloped bass response.  That's where some type of EQ comes into play with any dipole.  Do the math, a 2m diameter baffle only gets your Fequal down to 58hz.  This sloped response shows prominantly in every real world measurement I've seen with only one exception, and that is the ripole woofer/subwoofer alignment which acts like a mechanical Linkwitz transform for dipoles, but that's off topic.

The other way to flatten response beyond the limitations of your baffle is with a high Qts driver, which has a hump in its natural IB response down near Fs, thus offsetting dipole cancellation in that range.  This is part of Martin's strategy even with his big baffles.  Keep in mind that musical taste plays a big roll in what level of OB bass performance is satisfying.  His baffles would stretch almost across my room, so there's just no way a reliance on high Q drivers could work with the bass performance I demand.  Plus my setup requires long speaker cables, so the damping of my amplifiers provides little assistance to keep woofer cones under control, and I'd just end up with high Q woofers flopping around at low frequency.  This is another area where simulations will let you down, because they won't be able to simulate the unstable driver response of a high Q driver below 100hz, and manufacturer's don't provide response curves for high Q woofers that are meaningful or realistic below 100-200hz.

Here's my advice for what it's worth, since the number crunchers don't think I know what I'm talking about:

1. Make sure you understand how dipole bass cancellation works....The front wave travels directly to you, and the rear wave is forced by the baffle to travel further to get to you.  They start off equal in magnitude and directly out of phase, so they net to zero wherever they have travelled the same distance.  Their phase relationship when they reach the listening position is what determines how much bass gets cancelled.  Linkwitz shows us the math if you are interested.  This basic concept applies to all folded alignments, even U-baffles (although they present some other considerations).  Once you're able to visualize this effect, then you also start to "see" what the rear wave is doing on its way to you, along with the impact of reflections within the structure itself and then the room.

2.  Set your goals in terms of bass extension and SPL, along with a baffle size you can live with.  Your listening habits and musical tastes are key in setting your goals.  Then you can plug that info and driver Sd into DipoleSPLmax, and you can quickly back into what you need in the way of woofers.

3.  Woofer selection is going to determine what kind of fullranger you need.  If you need a lot of deep bass, then your woofers will be of a type that can't play very high, so you need a higher Q main driver which will reach down to the 100hz range.  This is where Martin's initial OB attempt broke down despite his "afternoon" optimizing simulations.  Those Dayton subwoofer type drivers simply weren't up to the task of blending with his Lowthers that start rolling off much higher than a typically recommended OB driver.  If you're set on using low Q fullrangers, then you have 2 choices, pro type woofers or multiple small woofers, both of which can be used higher in frequency, but you can forget about deep bass because the Fs is too high.


Martin,
I'm not trying to pick on you or your methods.  It's just that using your well documented project in my examples was the best way to get my points across.  I really intended this type of post for another time, but I didn't want to leave Florian hanging and no one else was stepping forward with something he could really use.  BTW,  I have several sets of OB's that I consider to be much more compromised.  In fact, I don't have anything that I would consider an all out effort and worthy of direct comparison, though I am getting closer.  Instead of running simulations and trying to quantify common designs, I've been working to optimize the physical baffle structure to take as much as possible into consideration.